skip to content
 

Clustering Sparse Graphs

Presented by: 
S Sanghavi University of Texas at Austin
Date: 
Wednesday 14th August 2013 - 10:00 to 10:45
Venue: 
INI Seminar Room 1
Abstract: 
Graph clustering involves the task of partitioning nodes, so that the edge density is higher within partitions as opposed to across partitions. A natural problem, it represents a first step in a wide array of applications in network analysis, community detection, recommendation systems etc.

A classic and popular statistical setting for evaluating between different solutions to this problem is the stochastic block model, also referred to as the planted partition model. In this talk, we present a new algorithm for this problem, which improves by polynomial factors over the performance of all previous known algorithms. It is based on convex optimization, and draws a connection between this problem and a different field: high-dimensional statistical inference.

The video for this talk should appear here if JavaScript is enabled.
If it doesn't, something may have gone wrong with our embedded player.
We'll get it fixed as soon as possible.
University of Cambridge Research Councils UK
    Clay Mathematics Institute The Leverhulme Trust London Mathematical Society Microsoft Research NM Rothschild and Sons